From AI Video to the No-Code Revolution: Understanding the Current State of Generative AI through the Top 100

In recent years, the Generative AI (GenAI) market has undergone remarkable evolution. In particular, chatbots utilizing Large Language Models (LLMs) and services capable of generating various media such as images, audio, and video have rapidly gained popularity, becoming familiar tools for general users.

Article:

In this article, based on discussions surrounding the A16Z (Andreessen Horowitz) report 'Top 100 GenAI Products, Ranked and Explained,' we will provide an easy-to-understand explanation of the overall trends in the GenAI market, emerging players, and the background behind the explosive increase in users. Additionally, we will explore future prospects by incorporating the features of each service and insights into their business models.

1. Overview and Background of the Top 100 List

1-1. Major Trends Indicated by the Top 100 List

The 'Top 100 GenAI Products' report, published by A16Z several times a year, compares and analyzes the number of users and active counts for both websites (desktop version) and mobile apps, presenting popular GenAI products in a ranking format. While ChatGPT, Midjourney, and Character AI, which have rapidly gained fame since around 2022, are prime examples, the rise of AI video generation services and products equipped with new natural language models is a key highlight this time.

1-2. Research Methodology: Web and Mobile Versions

The ranking is aggregated along two axes: the 'Top 50 Web Version' and the 'Top 50 Mobile App Version'.

• Web Version: Top sites are extracted based on SimilarWeb's monthly visit counts. However, the condition is that the service must be purely 'GenAI-centric,' excluding services where AI is merely a single functional feature.

• Mobile Version: Based on two types: Sensor Tower's monthly active users (MAU) and the newly introduced 'App Revenue (Sales) Ranking.' The former indicates the number of users, while the latter shows the top 50 in billing revenue; comparing the two reveals the differences between business models and user counts.

1-3. The Newly Established 'Brink List'

In this report, the 'Brink List,' which includes services that narrowly missed the top 50, was also published for the first time. This indicates a group of noteworthy services that have seen recent momentum in user numbers and are expected to rise in future rankings. It includes several all-in-one platform services that integrate Large Language Models (LLMs) and image generation AI, holding the potential to reshape future rankings.

2. Emerging Players and Key Trends

2-1. The Breakthrough of AI Video Generation

A highlight of this ranking is the rise of AI video generation services. In addition to the existing Sora (OpenAI), models from China such as Hyo and Clling have also broken into the top ranks. Chinese companies, in particular, are demonstrating international competitiveness due to their large number of researchers and the abundance of data used for image and video training. On the other hand, it has been pointed out that because copyright and licensing regulations are relatively looser compared to the West, they are capable of more realistic expressions.

The Potential of AI Video

• Cost Reduction Effect: The core concept of generative AI is to 'lower creation costs.' This is permeating the video field, and the generation range is beginning to expand experimentally from short clips to tens of seconds or even minutes.

• Combination of Multi-Models: Attempts to integrate multiple modalities such as 'video + subtitles + audio' are active. Beyond video generation via text prompts, users' creative freedom is increasing, allowing them to specify effects and camera angles.

2-2. Generalization of Vibe Coding (No-Code Generation)

In the realm of 'generating apps from text' and 'creating prototypes without code,' services like Bolt, Lovable, and Cursor have rapidly gained popularity and entered the rankings. These are also known as 'vibe coding,' which are platforms that allow users to easily create apps and websites without writing professional code.

• Expansion of the target user base: Initially perceived as tools for engineers, the participation of non-engineers is increasing due to simple UIs and automatic prompt optimization.

• The rise of DIY/personal software: Because personal ideas or spontaneous joke apps can be created and published in a few hours, the barrier to validating ideas has dropped significantly. The concept of "Disposable Software" has also emerged, making it easier to try out apps that are not intended for long-term operation.

3. The enduring popularity of standard services

3-1. ChatGPT and its ecosystem

OpenAI's ChatGPT consistently reigns at the top of the rankings. While it gained massive traffic from student users from the time of its initial release, it has recently expanded its user base again through extensions such as new voice modes and image recognition features (multimodal support), with reports indicating that active users have doubled in the last six months alone.

Meanwhile, assistant-type LLMs with established reputations for quality, such as Anthropic's Claude and the recently surging DeepSeek, are also increasing and gathering user support. DeepSeek, in particular, jumped to second place in the rankings shortly after its release, leaving the impression that "ChatGPT is not the only dominant player."

3-2. The rise of entertainment and communication AI

Character AI, which specializes in dialogue with AI characters, and services that generate interactive NSFW content also enjoy enduring popularity. These are characterized by users starting to use them to "enjoy fictional worldviews or their own unique stories," and combined with the enthusiasm of creative communities, they continue to see a steady influx of users.

4. The reality of GenAI from the perspective of revenue models

4-1. The gap between top revenue and top users

In this report, we extracted the "Top 50 Monthly Active Users (MAU)" and "Top 50 In-App Purchase Revenue" for mobile apps, and found that the overlap between the two rankings was only about 40%. In other words, "a large number of users does not necessarily mean high revenue."

• Apps that gather a wide and shallow user base: In many cases, they monetize through in-app ads or low-cost subscriptions. The number of users is huge, but the ARPU (Average Revenue Per User) is low.

• Apps that monetize narrowly and deeply: For example, in fields like education or photo/video editing, they set higher subscription fees for users who want to use the service seriously. Even if the number of users is small, they maintain a high level of revenue.

4-2. The balance between capital recovery and retention

Some app developers actively invest in acquiring paid users by utilizing ad slots on the App Store. It is not uncommon for indie developers or app studios to operate on a small scale rather than relying on investment from Venture Capital.
Furthermore, whether to make features partially free or to fully monetize them to aim for high ARPU results in significant differences in user growth and conversion rates. Such empirical knowledge will become increasingly important in the GenAI market, which is still in its infancy.

5. Future outlook: Premature hypothesis collapse and further chaos

5-1. The speed at which the "obvious" is overturned in months

In music and image generation AI, the prediction that "humans will not listen to or be able to distinguish AI-made works" collapsed within a few months, and with new models appearing—such as DeepSeek and Claude entering a market once thought to be dominated solely by ChatGPT—the old "common sense" is being broken one after another. This is because the cycle of "research paper -> model implementation -> app development -> consumer adoption" is shortening.

5-2. Deepening of Multimodal and Operator Capabilities

Multimodal AI, which handles images, video, audio, and text across domains, is expected to evolve further, with app integration and automated task execution (operator functions) becoming commonplace.
For example, models that can perform screen operations on behalf of the user, such as OpenAI's "Operator mode" and Google's "Gemini Flash," have emerged. If AI that functions as a "behind-the-scenes" assistant—handling tasks like payment processing and various administrative procedures—becomes widespread, it is highly likely that the user experience will change significantly.

What becomes clear through the A16Z ranking is the "speed" and "diversity" of GenAI. While emerging services that did not exist just a few months ago are gathering hundreds of thousands to millions of users and rising rapidly to the top, even famous apps that have once established a stable position can have their rankings threatened by the arrival of powerful new models.
To survive in this volatile market, the key will be not only technology but also user experience and the definition of clear use cases. Furthermore, as trends like "multimodal" and "decreasing generation costs" continue to progress, business opportunities are expected to expand far beyond traditional IT sectors, spreading into every industry, including music, video, education, and healthcare.

In this article, we have provided an overview of the latest trends in A16Z's "Top 100 GenAI Products" and the major players and trends observed there. Now that the cycle of research and business application is becoming increasingly short, there is even a possibility that the next report will feature a completely different lineup from this ranking. We cannot take our eyes off the future of the ever-evolving GenAI market.

Related Articles